To simulate dataset size heterogeneity, we follow a power law to randomly partition total {3000, 6000, 12000} MNIST examples among {5, 10, 20} participants ...
確定! 回上一頁